Understanding Operator Attitudes Toward AI-Supported Decision Making in Maritime Operations

Understanding Operator Attitudes Toward AI-Supported Decision Making in Maritime Operations (Jirak, Saroukanoff, van Rooy, arXiv:2609.11805, September 2026) reports survey results from 66 maritime professionals (including Captains and Officers of the Watch) regarding AI collision-avoidance assistants.

• Trust and Anxiety: Participants showed a generally positive disposition toward technology with no clear age-related differences in openness; however, trust remained moderately favorable but cautious, with stable trust levels across different traffic complexity scenarios.
• Perceived Advantages: Seafarers valued AI support for confidence-building (especially for junior officers), enhancing situation awareness, and time-saving for deeper analysis.
• Key Concerns: Critical risks identified included over-reliance, skill degradation, and input reliability issues (e.g., sensor failures or GPS spoofing), leading to calls for calibrated reliance rather than unconditional trust.
• Design Implications: The study suggests future systems must prioritize transparent, reliable, and operationally meaningful explanations (Explainable AI) to support human expertise and mitigate cognitive overload or alarm fatigue.

Summary: This arXiv preprint examines human factors in AI-supported decision making within maritime operations, with a focus on AI collision-avoidance assistants used by maritime professionals. Based on reported survey results, the material investigates how practitioners perceive and respond to AI decision support in a high-stakes, time-constrained operational setting. It centers on three interrelated constructs: trust in the AI system, technology anxiety, and perceived explanation quality. Rather than treating AI acceptance as a purely technical question, the work frames it as a human–automation interaction problem in which operator confidence, comfort with the technology, and the intelligibility of AI recommendations all shape whether the system is used effectively.

The key contribution is an empirical, domain-specific account of the factors that influence maritime operators’ willingness to rely on AI collision-avoidance support. The material suggests that trust is not determined only by accuracy or system performance, but is also shaped by how well the system explains its reasoning and how much anxiety operators experience when delegating or integrating AI-generated recommendations. Perceived explanation quality appears to be especially important in collision avoidance, where operators must interpret machine outputs under uncertainty, limited time, and potentially life-safety consequences. The findings imply that effective AI assistants in this domain should provide not only reliable recommendations, but also transparent, task-relevant explanations that help operators understand the basis of the advice, assess its urgency, and decide when to accept, modify, or override it.

This work matters because maritime operations are moving toward increasingly AI-assisted navigation, yet safe deployment depends on more than algorithmic performance. The material contributes to the broader literature on human–AI teaming by highlighting the role of trust calibration, explainability, and affective responses in a safety-critical operational environment. Its insights are relevant to system designers, maritime safety researchers, regulators, and operators involved in developing or deploying AI-based collision-avoidance tools. By identifying where operator concerns and explanation needs intersect, the study helps inform the design of AI assistants that are not only technically capable, but also credible, interpretable, and aligned with the practical realities of bridge decision making.